On the inverse scattering from anisotropic periodic layers and transmission eigenvalues

Dinh-Liem Nguyen; Isaac Harris; Jonathan Sands; Trung Truong

REVIEW 3 major objections 4 minor 21 references

On the inverse scattering from anisotropic periodic layers and transmission eigenvalues

T0 review · 3 major / 4 minor · reviewed 2026-08-14 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read This paper proves that a single series test over the eigensystem of the near-field operator decides membership in the support of an anisotropic periodic layer, and that the quasi-periodic transmission problem has infinitely many…

desk verdict Worth a careful referee, but the central factorization has a sign error that propagates into the main characterization; fixable, but as written the proof of Theorem 3.6 does not go through. read the letter →

arxiv 1908.05801 v2 pith:JPNY3EGU submitted 2019-08-16 math.AP

Isaac Harris , Dinh-Liem Nguyen , Jonathan Sands , Trung Truong This is my paper · ORCID

classification math.AP MSC 35R3078A4665C20

keywords inversescatteringfactorizationmethodtransmissioneigenvaluesanisotropicperiodicstructuresquasi-periodiclayersnear-fielddatashapereconstructionTMpolarization

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper claims two things about anisotropic periodic layers in two dimensions. First, the support of the layer's contrast can be recovered from near-field scattering data by a pointwise series test: a point belongs to the support exactly when a certain sum over the eigensystem of the data operator converges. Second, the associated quasi-periodic interior transmission eigenvalue problem has infinitely many real eigenvalues. Both results matter for periodic structures that model photonic crystals and gratings, where knowing the layer's shape from scattering data is a nondestructive-testing task and transmission eigenvalues serve as material signatures. The paper supplies proofs for both claims and numerical examples for the shape-reconstruction method.

What carries the argument

The argument turns on the factorization $WN = H^* T H$, where $N$ is the near-field operator, $W$ is an explicit block weight matrix with a bounded inverse, $H$ is a compact injective map from sequence data into the layer, and $T$ is a contrast operator defined by $T f = (\mathrm{Re}\,Q)^{-1/2} Q((\mathrm{Re}\,Q)^{-1/2} f + \nabla u)$. Under Assumption 3.1, $T$ is injective, has negative semidefinite imaginary part (strictly negative on nonzero vectors), and differs from a coercive operator by a compact perturbation; the range identity theorem then gives $\operatorname{Range}(H^*) = \operatorname{Range}((WN)_\sharp^{1/2})$. The quasi-periodic Green function provides test vectors $r_n(z)$ whose Rayleigh coefficients are known explicitly, so membership of $z$ in $D$ is read off from the finiteness of the eigensystem sum. For the eigenvalue half, the key object is $L_k$, defined by a Riesz representation from a Hermitian sesquilinear form; its self-adjointness, coercivity at $k=0$, and compactness of $L_k - L_0$ allow the standard counting argument for transmission eigenvalues, and a family of disjoint balls inside $D$ produces the infinite-dimensional subspaces on which $L_k$ is non-positive.

What would settle it

Take a non-absorbing anisotropic periodic layer ($\operatorname{Im} Q = 0$) and compute the near-field operator $N$; if the series in (19) converges at some point outside $D$ or diverges at some point inside $D$, the characterization fails. For the eigenvalue result, discretize $L_k$ for a layer with $A$ a constant positive-definite multiple of the identity: the theorem predicts infinitely many real eigenvalues, so observing only finitely many sign changes of $L_k$ would falsify Theorem 4.3.

Watch

Extended reading notes

Core claim

The central discovery is a complete characterization and a countable spectrum. Theorem 3.6 says that, under the absorbing-contrast Assumption 3.1, a point $z$ lies in the support $D$ exactly when the series $\sum_j |\langle r_n(z), \psi_{n,j}\rangle|_{\ell^2(\mathbb{Z})^2}^2 / \lambda_j$ is finite, where $(\lambda_j, \psi_{n,j})$ is an orthonormal eigensystem of the self-adjoint positive operator $(WN)_\sharp^{1/2}$ built from the measured near-field operator $N$. This is an if-and-only-if test, so it is both a uniqueness proof and a pixel-by-pixel imaging criterion. Theorem 4.3 states that, for a non-absorbing layer ($\operatorname{Im} Q = 0$) with uniformly positive contrast and quasi-periodicity parameter $\alpha$ not an integer, there exist infinitely many transmission eigenvalues. The proof uses an operator $L_k$ on $H^1_{0,\alpha}(D)$ that is self-adjoint, coercive at $k=0$, and compactly different from $L_0$; positivity for small $k$ and non-positivity on an $M_\epsilon$-dimensional subspace built from disjoint balls force the eigenvalue count to grow without bound. Together the two theorems give the first factorization-method justification and transmission-eigenvalue existence result for the anisotropic periodic-layer setting.

Load-bearing premise

The inverse-scattering result collapses if the contrast is not absorbing: the proof needs $\operatorname{Im} Q$ strictly negative definite so that $T$ is injective and the range identity theorem applies, and the paper explicitly notes this assumption excludes transmission eigenvalues.

Editorial extensions

If this is right

The series test is a yes/no criterion for every point, so the layer's support can be imaged pixel by pixel from the measured near-field operator without knowing the contrast's amplitude.
Because the criterion is an equivalence, the near-field operator determines the support $D$ uniquely among layers satisfying Assumption 3.1.
The proof requires near-field data taken from both sides of the layer, and the numerical examples show that evanescent modes are needed for resolution.
The infinite sequence of real transmission eigenvalues gives countable material-dependent fingerprints that are in principle recoverable from scattering data.
The absorbing assumption behind the factorization and the transparency assumption behind the eigenvalue result delimit two complementary regimes: one for imaging, one for material probing.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

Because the proof of Lemma 3.5 is deferred to the half-space analogue, the self-containedness of the full-space argument rests on trusting that the half-space estimates carry over; a reader should verify the sign and coercivity estimates for the full-space Dirichlet-to-Neumann maps directly.
The paper reports stable reconstructions for ball- and cross-type layers but noise-sensitive ones for piecewise-linear and sinusoidal layers, and offers no justification; this suggests the imaging functional's stability depends on layer topology, which is a testable hypothesis.
The factorization structure $WN = H^* T H$ appears portable to full Maxwell biperiodic problems, where $W$ would encode polarization-dependent Dirichlet-to-Neumann data; the paper does not attempt this.
The disjoint-ball counting argument for transmission eigenvalues hints that the number of eigenvalues grows with the volume of $D$; a Weyl-type asymptotic is a natural next step.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Colleague,

Here's my read on arXiv:1908.05801. The paper does two things: adapts the factorization method to anisotropic periodic layers in full space with two-sided near-field data, and proves existence of infinitely many transmission eigenvalues for the quasi-periodic transmission eigenvalue problem. The first is a genuine extension of Nguyen's half-space work [18] and avoids the connectivity assumption in T.-P. Nguyen [19]. The second follows the Cakoni–Kirsch machinery but the formulation with quasi-periodic conditions is new. The numerical examples show the method works when it works, though the authors admit the reconstructions for the first two geometries are unstable and the displayed figures are best-of-10.

Now the serious problem. The sign in the factorization is wrong. Let me walk through it. Equation (7) defines u via B(u,v;A) = -∫_D Q(ReQ)^-1/2 f·∇v. Since A=I+Q and Q is supported in D, B(u,v;A)=B(u,v;I)+∫_D Q∇u·∇v. Therefore B(u,v;I) = -∫_D Q((ReQ)^-1/2 f+∇u)·∇v. But T in (18) is defined so that (ReQ)^{1/2}Tf = Q((ReQ)^-1/2 f+∇u). Hence u solves B(u,v;I) = -∫_D (ReQ)^{1/2}Tf·∇v, while E in (15) is defined with the opposite sign: B(u,v;I)=+∫_D (ReQ)^{1/2}f·∇v. So G = -ET, not ET. Lemma 3.4's conclusion WN = H*TH should be WN = -H*TH. This is not cosmetic: Lemma 3.5(b) says ImT ≤ 0, so the correct factorization gives Im(WN) ≥ 0, and the range-identity functional in Theorem 3.6 should be |ReWN| + ImWN, not |ReWN| - ImWN. As written, the range identity from [14] does not apply. The error looks fixable by redefining T (or E) with a minus sign, but the proof of Theorem 3.6, as printed, doesn't hold.

Other soft spots are minor by comparison. Lemma 3.5 is stated without proof ("omitted here, following [18]") and Theorem 3.6's proof is "similar to [18]"; that's acceptable for a subfield that knows the machinery, but it makes the paper less self-contained. The numerical section is honest about instability, but showing only the best of 10 runs is not a strong demonstration for the two unstable shapes.

The transmission eigenvalue part looks more solid: the operator L_k is self-adjoint, L_0 is coercive, L_k-L_0 compact, and the dimension-counting argument with shrinking balls gives infinitely many eigenvalues. I didn't spot an error there.

Who is this for? Researchers working on factorization methods for periodic structures and on transmission eigenvalue problems for quasi-periodic settings. They should read it with a pencil. I'd send it to peer review—the sign error is real but likely fixable, and the contributions are worth having in the literature after correction. I would not cite it in its current form.

Referee Report

3 major / 4 minor

Summary. The paper studies two problems for a 2D anisotropic periodic layer in TM polarization: the inverse scattering problem of reconstructing the support D of the contrast from near-field data, and the interior transmission eigenvalue problem for non-absorbing anisotropic periodic layers. For the inverse problem, the authors propose a factorization method based on the two-sided near-field operator N and prove a characterization of D in terms of an eigensystem of a symmetrized operator (WN)_♯. They also provide numerical reconstructions for several layer geometries. For the transmission eigenvalue problem, they formulate a quasi-periodic interior transmission eigenvalue problem and prove the existence of infinitely many real transmission eigenvalues under the assumption Im Q = 0 and α ∉ Z.

Significance. If correct, the factorization result would provide a unique and computationally simple characterization of the shape of anisotropic periodic layers from near-field data, extending earlier half-space results to the full-space two-sided setting. The transmission eigenvalue existence result for quasi-periodic eigenfunctions is new and appears to be essentially sound. The paper is clearly written and the numerical experiments, despite the selection issue noted below, illustrate the method. However, the central factorization theorem in Section 3 contains a sign error that, as written, invalidates the range-identity argument, and key lemmas are stated without proof. The transmission eigenvalue part of the paper is the stronger contribution; the inverse part needs substantial correction before the claims can be accepted.

major comments (3)

[§3.2, Lemma 3.4] There is a sign error in the factorization. From (7), B_A(u,v) = -∫_D Q(ReQ)^{-1/2} f·∇v. Since A = I + Q, B_A(u,v) = B_I(u,v) + ∫_D Q∇u·∇v, and using (18) we obtain B_I(u,v) = -∫_D (ReQ)^{1/2} Tf·∇v. Therefore the unique solution of (7) satisfies Gf = -E Tf, not Gf = E Tf as claimed in the proof of Lemma 3.4. Consequently Lemma 3.4 should read WN = -H^* T H. This is load-bearing: with Im T ≤ 0 from Lemma 3.5(b), the identity WN = -H^* T H gives Im(WN) = -H^*(Im T)H ≥ 0, so the operator (WN)_♯ = |Re WN| - Im WN used in Theorem 3.6 need not be positive definite, and the range identity theorem [14] cannot be applied as written. The proof of Lemma 3.4 drops the minus sign when rewriting (7) as an identity for B_I; this must be corrected, and the definition of ♯ or T must be adjusted accordingly.
[§3.1, Eq. (17)] The Rayleigh coefficients of the quasi-periodic Green's function (16) are not those given in (17). For x2 > h, the coefficient of e^{iα_n x1 + iβ_n(x2-h)} in G is r+_n(z) = i/(4πβ_n) e^{-iα_n z1} e^{iβ_n(h-z2)}, and for x2 < -h it is r-_n(z) = i/(4πβ_n) e^{-iα_n z1} e^{iβ_n(z2+h)}. Equation (17) has the opposite sign in the z2-dependent phase, i.e., it gives the coefficients of the incoming (conjugate) Green's function. Since Lemma 3.3 and the test functional in Theorem 3.6 both use r_n(z), the characterization of D is not justified unless this formula is corrected and the proof of Lemma 3.3 is checked with the correct phase.
[§3.2, Lemma 3.5 and Theorem 3.6] These statements are load-bearing for the central inverse-scattering claim, yet they are not proved in the manuscript. Lemma 3.5 states that the proofs 'are omitted here, following [18]', and Theorem 3.6 states that 'the proof is similar to [18]'. The full-space two-sided problem has a different measurement operator, a different weighting operator W, and different boundary terms than the half-space problem in [18]; the sign issues identified above show that the adaptation is not automatic. Please provide complete proofs of Lemma 3.5 and Theorem 3.6, or a detailed step-by-step mapping to [18] that accounts for all sign conventions and the two-sided nature of the data.

minor comments (4)

[§3.1, Eq. (14)] The matrix W in (14) does not match the formula for H^* in (11). To have H^* = W E, the first row of W should be ( ~w+_n, ~w+_n ), not ( ~w+_n, ~w-_n ). The determinant and invertibility argument are unaffected.
[§3.1, Lemma 3.3] In the proof of the z ∈ D direction, the displayed formula should read (Ef)_n = ( \hatΦ^+_n, \hatΦ^-_n )^T rather than ( \hatΦ^+_n, \hatΦ^+_n )^T.
[§5, Numerical examples] Figures 2(d) and 3(d) are described as the best results out of 10 numerical experiments, and the worst cases are stated to show no reasonable reconstruction. As presented, this does not give a robust measure of the method's performance. Please report median or typical reconstructions and show at least one representative non-best case, or explain the selection criterion.
[Abstract and §6] There are minor language and typographical issues, e.g., 'transmission eigenvalues in scattering have recently attracted' in the abstract and 'Lipchitz' in Section 6; these should be corrected.

Circularity Check

0 steps flagged · score 2.0 of 10

No circular reduction; central results rest on external theorems, with a minor self-citation for omitted half-space analogues.

full rationale

Walking the derivation chain: the factorization method in Theorem 3.6 is assembled from Lemma 3.3 (characterization through Range(H*)), Lemma 3.4 (factorization WN = H*TH), the external range identity theorem [14], and the analytic properties of H and T in Lemma 3.5. The adjoint H* is computed explicitly, Lemma 3.3 has a direct proof, and Lemma 3.4 derives the factorization from N = GH and H* = WE rather than postulating it. Lemma 3.5 and Theorem 3.6 are explicitly deferred to [18] ('The proofs for these properties can be done following their analogues of the half-space case [18] and therefore are omitted here'; 'The proof is similar to that of the half space case [18]'). This is a self-citation by a coauthor, and it is load-bearing in the sense that central details are not written out. However, [18] is a published half-space analogue with different geometry and assumptions, not the target full-space periodic-layer result, so the citation supplies independent support rather than a circular reduction. The transmission eigenvalue result in Theorem 4.3 uses the standard Cakoni-Kirsch theory [8] and constructs non-positivity of L_k from the smallest transmission eigenvalue of a constant-coefficient ball problem; that auxiliary existence is a known external fact, not the theorem being proved. No parameter is fitted and then renamed as a prediction, and no step reduces to an input by definition. The numerical section uses the authors' own forward solver [17] to generate synthetic data, but this is a consistency demonstration rather than a derivation. The score of 2 reflects only the minor self-citation for omitted proof details; no circularity is present. Any alleged sign inconsistency in the factorization would be a correctness issue, not a circularity issue, and does not affect this assessment.

Assumptions & free parameters 1 free parameters · 8 assumptions · 0 invented entities

The mathematical derivations are parameter-free: no constants are fitted to data in the proofs. Numerical examples use k=5.85, M=10/20, 5% noise, and a truncation threshold, all inputs rather than fitted parameters. The axioms are standard spectral theory, the range identity theorem [14], and the ball transmission eigenvalue existence theory [8], plus explicit domain assumptions stated in Assumption 3.1 and Section 4. No new physical entities are introduced.

free parameters (1)

Singular value truncation threshold = 5e-4
Used in Section 5 to regularize noisy data by dropping singular values of (WN)_sharp below 5e-4. It is a hand-chosen numerical regularization parameter and does not enter the mathematical theorems.

assumptions (8)

standard math Uniform ellipticity of the matrix A and well-posedness of the direct scattering problem for all but a discrete set of wave numbers k.
Invoked in Section 2; the paper cites [4] for well-posedness and assumes k is such that the direct problem is uniquely solvable.
domain assumption Assumption 3.1: D is a Lipschitz domain, the contrast Q is complex symmetric with Re Q positive definite and Im Q negative definite, and (Re Q)^{+/-1/2} are bounded.
This is the core hypothesis for the factorization method, making the operator T injective with sign-definite imaginary part (Lemma 3.5).
domain assumption The complement Omega \ D has at most two unbounded connected components (Assumption 3.1).
Needed for the analytic continuation argument in Lemma 3.3 that proves the range characterization of D.
domain assumption alpha not in Z so that non-zero constants are not quasi-periodic and the Poincare inequality holds for H1_alpha(D) (Lemma 4.1).
Required for the operator L0 to be coercive and for Lax-Milgram in equation (27).
domain assumption In Section 4 the scatterer is non-absorbing (Im Q = 0) with Q uniformly positive definite, and D is a graph-type Lipschitz domain with no holes touching the vertical boundaries.
The transmission eigenvalue problem is formulated for this geometry; the summary states the hole-free and Lipschitz restriction.
standard math The range identity theorem from Kirsch-Grinberg [14].
Used to pass from Range(H*) to Range((WN)_sharp^{1/2}) in Theorem 3.6.
standard math The existence of infinitely many transmission eigenvalues for the constant-coefficient ball problem (30)-(31), from Cakoni-Kirsch [8].
Used in Theorem 4.3 to construct the functions u_epsilon that make L_k non-positive.
standard math Rellich compact embedding and the spectral theorem for compact self-adjoint operators.
Used in Lemma 4.1, Lemma 4.2, and the Picard series criterion (19).

how reviews work

0 comments

Cite this review

Pith. "Pith review of On the inverse scattering from anisotropic periodic layers and transmission eigenvalues." pith.science (2026). https://pith.science/paper/JPNY3EGU

@misc{pith2026190805801,
  author       = {Pith},
  title        = {Pith review of: On the inverse scattering from anisotropic periodic layers and transmission eigenvalues},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/JPNY3EGU}},
  note         = {Machine review of arXiv:1908.05801}
}

read the original abstract

This paper is concerned with the inverse scattering and the transmission eigenvalues for anisotropic periodic layers. For the inverse scattering problem, we study the Factorization method for shape reconstruction of the periodic layers from near field scattering data. This method provides a fast numerical algorithm as well as a unique determination for the shape reconstruction of the scatterer. We present a rigorous justification and numerical examples for the factorization method. The transmission eigenvalue problem in scattering have recently attracted a lot of attentions. Transmission eigenvalues can be determined from scattering data and they can provide information about the material parameters of the scatterers. In this paper we formulate the interior transmission eigenvalue problem and prove the existence of infinitely many transmission eigenvalues for the scattering from anisotropic periodic layers.

Figures

Figures reproduced from arXiv: 1908.05801 by the authors.

**Figure 2.** Shape reconstruction for the periodic layer of piecewise linear type. [PITH_FULL_IMAGE:figures/full_fig_p018_2.png] view at source ↗

**Figure 3.** Shape reconstruction for the periodic layer of sinusoidal type. [PITH_FULL_IMAGE:figures/full_fig_p019_3.png] view at source ↗

Figures from the paper (2 more)

**Figure 4.** Figure 4: Shape reconstruction for the periodic layer of ball type. [PITH_FULL_IMAGE:figures/full_fig_p019_4.png]

**Figure 5.** Figure 5: Shape reconstruction for the periodic layer of cross type. [PITH_FULL_IMAGE:figures/full_fig_p019_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

21 extracted references · 21 canonical work pages

[18]

Nguyen , Shape identiﬁcation of anisotropic diﬀraction gratings for TM- polarized electromagnetic waves, Appl

D.-L. Nguyen , Shape identiﬁcation of anisotropic diﬀraction gratings for TM- polarized electromagnetic waves, Appl. Anal., 93 (2014), pp. 1458–1476

work page 2014
[19]

Nguyen , Diﬀerential imaging of local perturbations in anisotropic periodic media, Inverse Problems, accepted (2019)

T.-P. Nguyen , Diﬀerential imaging of local perturbations in anisotropic periodic media, Inverse Problems, accepted (2019)

work page 2019
[14]

Kirsch and N

A. Kirsch and N. Grinberg , The Factorization Method for Inverse Problems , Oxford Lecture Series in Mathematics and its Applications 36, Oxford University Press, 2008

work page 2008
[1]

Arens and N

T. Arens and N. Grinberg , A complete factorization method for scattering by periodic structures, Computing, 75 (2005), pp. 111–132

work page 2005
[2]

Arens and A

T. Arens and A. Kirsch , The factorization method in inverse scattering from periodic structures, Inverse Problems, 19 (2003), pp. 1195–1211

work page 2003
[3]

G. Bao, T. Cui, and P. Li , Inverse diﬀraction grating of Maxwell’s equations in biperiodic structures, Optics Express, 22 (2014), pp. 4799–4816

work page 2014
[4]

Bonnet-Bendhia and F

A.-S. Bonnet-Bendhia and F. Starling , Guided waves by electromagnetic gratings and non-uniqueness examples for the diﬀraction problem , Math. Meth. Appl. Sci., 17 (1994), pp. 305–338

work page 1994
[5]

Cakoni, D

F. Cakoni, D. Colton, and H. Haddar , On the determination of dirichlet or transmission eigenvalues from far ﬁeld data , C. R. Acad. Sci. Paris, 348 (2010), pp. 379–383

work page 2010

Show all 21 references

[6]

, Inverse Scattering Theory and Transmission Eigenvalues , SIAM, 2016

2016
[7]

Cakoni, H

F. Cakoni, H. Haddar, and I. Harris , Homogenization of the transmission eigenvalue problem for periodic media and application to the inverse problem , In- verse Probl. Imaging, 9 (2015), pp. 1025–1049

2015
[8]

Cakoni and A

F. Cakoni and A. Kirsch , On the interior transmission eigenvalue problem , Int. Jour. Comp. Sci. Math., 3 (2010), pp. 142–167

2010
[9]

Elschner and G

J. Elschner and G. Hu , An optimization method in inverse elastic scatter- ing for one-dimensional grating proﬁles , Commun. Comput. Phys., 12 (2012), pp. 1434–1460

2012
[10]

Haddar and T.-P

H. Haddar and T.-P. Nguyen , Sampling methods for reconstructing the ge- ometry of a local perturbation in unknown periodic layers , Comput. Math. Appl., 74 (2017), pp. 2831–2855

2017
[11]

Harris, F

I. Harris, F. Cakoni, and J. Sun , Transmission eigenvalues and non- destructive testing of anisotropic magnetic materials with voids , Inverse Problems, 30 (2014), p. 035016

2014
[12]

Kleefeld and L

A. Kleefeld and L. Pieronek , Computing interior transmission eigenvalues for homogeneous and anisotropic media , Inverse Problems, 34 (2018), p. 105007

2018
[13]

Jiang and P

X. Jiang and P. Li , Inverse electromagnetic diﬀraction by biperiodic dielectric gratings, Inverse Problems, 33 (2017), p. 085004. 20

2017
[15]

Kirsch and A

A. Kirsch and A. Lechleiter , The inside-outside duality for scattering prob- lems by inhomogeneous media , Inverse Problems, 29 (2013), p. 104011

2013
[16]

Lechleiter and D.-L

A. Lechleiter and D.-L. Nguyen , Factorization method for electromagnetic inverse scattering from biperiodic structures , SIAM J. Imaging Sci., 6 (2013), pp. 1111–1139

2013
[17]

, A trigonometric Galerkin method for volume integral equations arising in TM grating scattering, Adv. Comput. Math., 40 (2014), pp. 1–25

2014
[20]

Sandfort, The factorization method for inverse scattering from periodic in- homogeneous media, PhD thesis, Karlsruher Institut f¨ ur Technologie, 2010

K. Sandfort, The factorization method for inverse scattering from periodic in- homogeneous media, PhD thesis, Karlsruher Institut f¨ ur Technologie, 2010

2010
[21]

J. Yang, B. Zhang, and R. Zhang , A sampling method for the inverse trans- mission problem for periodic media , Inverse Problems, 28 (2012), p. 035004. 21

2012

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.

[1] [18]

Nguyen , Shape identiﬁcation of anisotropic diﬀraction gratings for TM- polarized electromagnetic waves, Appl

D.-L. Nguyen , Shape identiﬁcation of anisotropic diﬀraction gratings for TM- polarized electromagnetic waves, Appl. Anal., 93 (2014), pp. 1458–1476

work page 2014

[2] [19]

Nguyen , Diﬀerential imaging of local perturbations in anisotropic periodic media, Inverse Problems, accepted (2019)

T.-P. Nguyen , Diﬀerential imaging of local perturbations in anisotropic periodic media, Inverse Problems, accepted (2019)

work page 2019

[3] [14]

Kirsch and N

A. Kirsch and N. Grinberg , The Factorization Method for Inverse Problems , Oxford Lecture Series in Mathematics and its Applications 36, Oxford University Press, 2008

work page 2008

[4] [1]

Arens and N

T. Arens and N. Grinberg , A complete factorization method for scattering by periodic structures, Computing, 75 (2005), pp. 111–132

work page 2005

[5] [2]

Arens and A

T. Arens and A. Kirsch , The factorization method in inverse scattering from periodic structures, Inverse Problems, 19 (2003), pp. 1195–1211

work page 2003

[6] [3]

G. Bao, T. Cui, and P. Li , Inverse diﬀraction grating of Maxwell’s equations in biperiodic structures, Optics Express, 22 (2014), pp. 4799–4816

work page 2014

[7] [4]

Bonnet-Bendhia and F

A.-S. Bonnet-Bendhia and F. Starling , Guided waves by electromagnetic gratings and non-uniqueness examples for the diﬀraction problem , Math. Meth. Appl. Sci., 17 (1994), pp. 305–338

work page 1994

[8] [5]

Cakoni, D

F. Cakoni, D. Colton, and H. Haddar , On the determination of dirichlet or transmission eigenvalues from far ﬁeld data , C. R. Acad. Sci. Paris, 348 (2010), pp. 379–383

work page 2010